Upload report_claigneau_inference_image-generation_onPremise.json

#48
data/stable_diffusion_berthelot/reports/report_claigneau_inference_image-generation_onPremise.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "licensing": "CC BY 4.0",
4
+ "reportDatetime": "2023-12-14 00:00:00",
5
+ "reportStatus": "draft",
6
+ "publisher": {
7
+ "projectName": "Adrien Berthelot, Eddy Caron, Mathilde Jay, Laurent Lef\u00e8vre. Estimating the environmental impact of Generative-AI services using an LCA-based methodology. 2023. \u27e8hal-04346102v1\u27e9",
8
+ "confidentialityLevel": "public"
9
+ }
10
+ },
11
+ "task": {
12
+ "taskStage": "inference",
13
+ "taskFamily": "image generation",
14
+ "nbRequest": 1,
15
+ "algorithms": [
16
+ {
17
+ "algorithmType": "Diffusion-based text-to-image generation model",
18
+ "foundationModelName": "Stable Diffusion",
19
+ "foundationModelUri": "https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5"
20
+ }
21
+ ],
22
+ "dataset": [
23
+ {
24
+ "dataUsage": "input",
25
+ "dataType": "word"
26
+ },
27
+ {
28
+ "dataUsage": "output",
29
+ "dataType": "image",
30
+ "dataQuantity": 4
31
+ }
32
+ ],
33
+ "taskDescription": "Input is one prompt but we don't know the size. 4 images have been generated"
34
+ },
35
+ "measures": [
36
+ {
37
+ "measurementMethod": "wattmeter",
38
+ "manufacturer": "Omegawatt",
39
+ "powerConsumption": 0.00138
40
+ }
41
+ ],
42
+ "infrastructure": {
43
+ "infraType": "onPremise",
44
+ "components": [
45
+ {
46
+ "componentName": "8 x Nvidia A100-SXM4-40GB GPUs",
47
+ "componentType": "gpu",
48
+ "nbComponent": 8,
49
+ "memorySize": 40,
50
+ "manufacturer_infra": "NVIDIA",
51
+ "series": "A100"
52
+ },
53
+ {
54
+ "componentName": "2 AMD EPYC 7742 CPUs",
55
+ "componentType": "cpu",
56
+ "nbComponent": 2,
57
+ "manufacturer_infra": "AMD",
58
+ "series": "EPYC 7742"
59
+ },
60
+ {
61
+ "componentName": "1 TiB RAM",
62
+ "componentType": "ram",
63
+ "nbComponent": 1,
64
+ "memorySize": 1024,
65
+ "manufacturer_infra": "Generic"
66
+ }
67
+ ]
68
+ },
69
+ "quality": "high"
70
+ }
data/stable_diffusion_berthelot/source_info.md ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Source : "Estimating the environmental impact of Generative-AI services using an LCA-based methodology
2
+
3
+ Source url : https://inria.hal.science/hal-04346102v2/document
4
+
5
+ Original authors : Adrien Berthelot, Eddy Caron, Mathilde Jay, Laurent Lefèvre
6
+
7
+ Original Publication date : 2023
8
+
9
+ Conversion authors : Clément Laigneau, Sarah Oury
10
+
11
+ Missing data : No info on the software stack used. No info on the size of text inputs and the resolution of the outputs. No info on the configuration of Stable diffusion
12
+
13
+ Model setback : None.